AI in Healthcare: How Artificial Intelligence Is Redefining Interoperability

For years, FHIR has been seen as a beacon of hope for the interoperability of medical data, but in practice, the standard is still in its infancy. Artificial intelligence (AI) could take over the conversion between different data formats in the future, thereby replacing standards—at least to a certain extent.

Why FHIR Has Only Worked in Isolated Cases So Far

In theory, FHIR enables the seamless exchange of patient data between practice management systems (PVS), hospital information systems (HIS), and electronic health records (eHR). In practice, however, the picture is different: Despite pilot projects, initiatives, and intensive discussions, its use in German hospitals has so far been fragmented.

This is primarily due to the high level of structuring that FHIR requires. Medical information is often generated in an unstructured form, such as in consultation notes, free-form assessments, or handwritten findings. Medical decisions are based on nuances and context that are difficult to squeeze into predefined data fields.

The documentation burden therefore increases with FHIR—and this against the backdrop that it already consumes too large a portion of working hours: According to the German Hospital Institute, physicians spend an average of three hours a day on formal documentation. New standardization requirements are therefore often viewed with skepticism, especially in complex, historically evolved hospital information system (HIS) landscapes.

AI as an Intelligent Translator

This is where the paradigm shift comes in: Instead of strictly standardizing data from the outset, it can first be captured in an unstructured format and then automatically converted into the required formats. Modern AI systems, particularly large language models (LLMs), make exactly that possible.

AI analyzes unstructured medical information, identifies correlations, and converts it into virtually any desired target format—including FHIR profiles. APIs remain relevant, but the complex logic of format conversion is shifting to intelligent AI layers that operate in the background. One could say that AI does not translate between languages, but between data worlds. The more widely AI is used in hospitals, the more likely this scenario becomes.

Practical Benefits for Hospitals and Medical Practices

The potential applications of AI in clinical practice are diverse and offer concrete benefits for staff and patients. For example, data from a practice management system can be transferred directly into the hospital information system without requiring additional manual work. Scanned documents from paper files can be automatically integrated into digital systems in a structured manner, making information available across systems at any time.

There is also great potential for the electronic health record (eHR): Patient data can be seamlessly consolidated regardless of its source, providing a complete picture of the patient’s medical history. At the same time, AI significantly reduces the documentation burden on physicians, leaving them with more time for direct patient care.

Increased Efficiency Through AI Workflows

Example of an AI-powered workflow: A physician documents a patient consultation with the help of an AI assistant that records the conversation and automatically structures the content. After a brief review, a second model generates FHIR-compliant resources or formatted medical reports, which are transferred directly to the hospital information system (HIS) or practice management system (PMS).

FHIR Remains Relevant—But in a Different Way

FHIR will continue to be needed—especially for data that is already structured, for research, registries, or billing. The key change is that AI serves as a flexible bridge between existing standards.

This makes interoperability in healthcare realistic, practical, and efficient: Hospitals benefit from seamless data integration without rigid standardization requirements burdening day-to-day clinical practice.